todo

[a-maumau]

✅add user information at terminal format

• from v0.5.3

✅add left ellipsis

• from nvidia-smi, long command seems to be already ellipsis at left.
• from v0.5.3

add more flexible setting/configuration

• currently, it always need to be rebooting.
• add theme or color configuration.
• we can trace the file date to detect update on setting file.

✅ add no log recorded mode (partially implemented -> add interval time for storing data.)

• for high frequency logging, it might be a bottleneck and large storage space will be consumed.

• we actually don't need the logs (?).

• or take some interval for recording.

  added an interval and it enable to force no recording. (v0.5.1 >=)

✅fix websocket reconnection

• it seems to be not working at current implementation.

• need to be watched the state of websocket object.

  fixed ٩( 'ω' )و✨

fix using magic number.

✅ fix typo in LICENSE
fixed ٩( 'ω' )و✨

[a-maumau]

✅bug found at gpu_info_sender.py, fix import to vesta.send_gpu_info.

• fixed ٩( 'ω' )و✨

[a-maumau]

✅BUG: in format_str.align_str(), suffix filling is not correct length.

• fixed from v0.4.2

✅CSS: style of last GPU entry's top margin doesn't exist.

• fixed from v0.4.2

[a-maumau]

slack bot is not reconnecting?

• for long time running, it sometimes lost the connection.
• python-slackclient 's bug or may my use is wrong.
  fixed from v0.5.3✨

it seems to be remained.
it depends on slackclient and it is v2.5.0 now, I hope it fill be fixed.

[a-maumau]

✅add Bot to reconnect to network in temporary down.

• from v1.0.0

it seems to be remain.
it depends on slackclient, so I hope this is fixed in slackclient v2.x

[a-maumau]

pie charts of each host's GPU memory usage which update from down to up don't show up correctly.

[a-maumau]

add feature : client down detecting for main page.

• server::http_server must be refactored to implement it.

✅bug: command ellipsis is not working right in printing process.

• fixed from v1.0.1

✅ to be fixed: notification occurs many time.

• reason is unknown (server or client).
• need to be queue style, wait for previous data to be recorded.

fix from v1.2.5, add exclusive controlling to avoid duplicated notification

[a-maumau]

to be fixed:
✅version is not correct

bug:
✅text bug in http://server_address/?term=true when it's down state

+------------------------------------------------------------------------------+
| vesta ver. 1.0.0                                                   gpu info. |
+------------------+------------------------+-----------------+--------+-------+
| host             | gpu                    | memory usage    | volat. | temp. |
+------------------+------------------------+-----------------+--------+-------+
| status: down                                                                 |
|   server1                 (***.***.***.***) last update: 1560/15/70 57:.8:11 |
+------------------+------------------------+-----------------+--------+-------+

last recorded data was around "2019/06/10 17:56:57.309"

• fixed from v1.1.0

[a-maumau]

to be fixed

• ✅Update slack client to 2.1.0
  • bot reconnection problem will be solved?
  • from v1.1.0, it is now slackclient v2.5.0
• Update other package too.

[a-maumau]

• ✅add new command for ps to read information
  • nvidia-smi's displaying has changed from around v.450

Fixed from v1.2.4

[a-maumau]

bug: duplicated registering

• using crontab, sometimes it register multiple times and get <host_name>_<hash_code> as a result.

[a-maumau]

possible bug

• "ImportError: sys.meta_path is None, Python is likely shutting down"
  • sometimes occur, possibly in gevent's cython, so need to update library version